The long-term objective of this research is to determine whether a simplified form of cardiopulmonary resuscitation instruction (involving the use of chest compressions only) provided by the dispatcher to the person calling to report a potential cardiac arrest, provides greater survival rates than ABC instructions involving airway, breathing and chest compression. A positive answer would: 1) prove that bystander CPR directly effects survival rates thus resolving the question of whether or not the observed association between bystander CPR and improved survival rates is the consequence of a common association with same third variable. 2) It would reduce the burden on the dispatcher by simplifying and shortening the interaction with the caller while providing better outcome. And, 3) it could have public health implications vis-a-vis the concern about transmission of AIDS. The study would be a randomized trial, with enrollment extending over 6.5 years (1.75 years will be completed by the date of funding for this application) conducted as follows: Each call requesting emergency medical aid for what the dispatcher diagnoses as cardiac arrest and for which CPR is not in progress or about to be initiated would represent an episode for randomization in the trial. A single key stroke on a microcomputer attached to each dispatcher's station results in randomization of that episode either to instructions involving chest compression only or to instructions involving the traditional airway, breathing and chest compression instruction for CPR. The microcomputer is programmed to facilitate the dispatcher with regard to the specific instruction set (a process that has been tested now for over five years). Subsequently, the paramedic field reports are reviewed to determine those patients who were actually in cardiac arrest. Study patients consist of those patients who were in cardiac arrest and whose episodes were randomized. The outcome measure is survival-defined as discharged alive from the hospital without serious neurological comorbidity. However, because survival from cardiac arrest for patients not found in ventricular fibrillation has been universally reported to be very poor, a key secondary analysis will compare survival in patients who are found in ventricular fibrillation. The primary analysis will compare survival in all randomized study patients.